Menü
Break the technical design into an ordered list of atomic, reviewable implementation tasks.
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Basierend auf der SOC-Berufsklassifikation
Dual-model adversarial review: Claude and Copilot review independently, verdicts are synthesized.
Implement the SDD task list. Read tasks and design from memory, write code, run tests.
Archive all SDD artifacts for a change as a single completed record in memory.
Write the technical design for a planned change: components, interfaces, data flows, edge cases.
Explore the codebase to build context for a planned change. First phase of the SDD workflow.
Propose 2–3 solution approaches for a planned change, evaluating trade-offs and recommending one.
| name | sdd-tasks |
| description | Break the technical design into an ordered list of atomic, reviewable implementation tasks. |
| when_to_use | After sdd-design. Reads design + spec from memory, produces a task list. |
| allowed-tools | ["Read","mcp__invencible-memory__memory_recall","mcp__invencible-memory__memory_save"] |
| model | sonnet |
| effort | low |
| user-invocable | false |
Decompose the design into the smallest independent tasks that can each be implemented and reviewed in a single focused session.
memory_recall(topic_key="sdd-<change_id>-design")memory_recall(topic_key="sdd-<change_id>-spec")SDD Tasks: <change_id>sdd-<change_id>-taskscontext## Task List
### T1: <title>
Files: <file paths>
What: <what to change>
Done when: <acceptance condition>
### T2: <title>
Files: <file paths>
What: <what to change>
Done when: <acceptance condition>
...
## Order Rationale
<why tasks are sequenced this way>